API Endpoint URLs
API Base URLs and Jev /v1/decisions or /v1/systemone requests: native payloads, incompatible chat endpoints, and convert_request_failed troubleshooting.
This page provides a centralized list of regional API endpoints and links for account and quota management.
1. Regional API Endpoints
| Region | Recommended Base URL | Description |
|---|---|---|
| Global (including Mainland China) | https://api.llm-token.cn/v1 | Globally routed: nodes in East China, South China, Hong Kong, Singapore, the US, and Germany; automatically picks the fastest route. |
2. Account and Quota Links
- Account Center: Add funds, renew service, and manage your account.
- Quota Lookup: Review your quota, usage, and credit status.
Jev native decisions: use a different URL, not chat
jev-latest uses the native decisions protocol, not an ordinary chat client. The gateway URL differs from TypeSafe's direct endpoint; the API keys are not interchangeable.
| Connection | Full POST URL | API key |
|---|---|---|
| Recommended gateway entry | https://api.llm-token.cn/v1/decisions | LLM API Gateway key |
| Compatible gateway entry | https://api.llm-token.cn/v1/systemone | LLM API Gateway key |
| Direct TypeSafe API | https://api.typesafe.ai/v1/systemone | TypeSafe key |
If the client's Base URL already ends in https://api.llm-token.cn/v1, append only /decisions or /systemone, not /v1/v1/decisions.
Include model, state, and questions. questions is an object keyed by question ID, with type and instructions for each question. For choice, criteria is an options object; for score, it is an ordered array; noul evaluates a yes/no probability. Do not send a chat messages payload or enable streaming.
Minimal request
Set your gateway API key in the LLM_API_KEY environment variable first. This example makes a real call; confirm your quota and billing before running it. Read answers.is_urgent.noul and usage, not choices[0].message.content.
curl -sS -X POST https://api.llm-token.cn/v1/decisions \
-H "Authorization: Bearer $LLM_API_KEY" \
-H "Content-Type: application/json" \
-d @- <<'JSON'
{
"model": "jev-latest",
"state": "The order is due today and the payment has failed.",
"questions": {
"is_urgent": {
"type": "noul",
"instructions": "Does this require urgent attention?"
}
}
}
JSONIf a 500 / convert_request_failed error contains this channel only supports native decisions requests, inspect whether the client still sends to /v1/chat/completions or /v1/messages. This is a protocol mismatch, not insufficient credit. Adding funds or choosing usage-based billing does not change the client's protocol. Correct the URL and payload before retrying; do not repeatedly resend the original chat request.